-—.

ao
03

FOR AERONAUTICS

TECHNICAL NOTE 2561

A STUDY OF POISSON’S RATIO IN THE YIELD REGION

By George Gerard and Sorrel Wildhorn “

New York University

. .. . . . . . . . . .. . . . .. . .. ..— .. d
IN NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS
icislzii‘
0ffb5Sbq

TECHNICALNOTE2563. ,

A STUDY OF POISSON’S RATIO IN TEE YIELD REGION

By George Gera~d and Sorrel Wildh.orm

sumARY

~ the yield region of the stress-strain curve the variation in

Poisson’s ratio from the elastic to the plastic value is most pronounced.
This variation waa studied experimentally by a systematic series of tests
on several aluminum alloys. The tests were conducted under simple tensile
and compressive loading along three orthogonal axes.

A theoretical variation of Poisson’s ratio for an orthotropic solid

was obtained from dilatational considerations. The assumptions used in
deriving the theory were examined by use of the test-data and were found
to be in reasonable agreement with experimental evidence.

INTRODUCTION
i.

Poisson’s ratio for engineering materials under s@le =i~ Zoa@

has a value in the elastic region of between 1/4 and 1/3 and On
usual.I.y
the assumption of a plastically incompressible isotropic solid assumes
a value of 1/2 in the plastic region. The transition from the elastic
to the plastic value, in general, is gradual and is most pronounced in
the yield region of the stress-strain cur=.

In the deformation theory of smll elmtic.and plastic strains for

an isotropic solid, which is summarized by Nadai in reference 1, it is
shown that the stress-strain relations for a strain-hardening material
depend essentially upon two deformation functions, the secant modulus
and the generalized Poisson’s ratio. Becaus’eof the fundamental nature
of the latter in any plasticity theory, this investigation w’s undertaken
to provide basic experimental data on the variation of Poisson’s ratio
b the yield region of some materials commonly employed in aircraft
applications.

General,dilatational relations are considered in the section entitled

“Theoretical Investigation” and it’is found that a theoretical relation-
ship for the variation of Poisson’s ratio from the elastic to the plastic
value can he obtained for an orthotropic medium in which the plane con-
taining the two isotropic axes is norml’to the applied load. This .

\,

, .-. - .... . -- .- -— --—-- _- —____ ,._. — —.— -.—. ———-— --- —— —-— --—-—--—---
 \
2 MCA TN 2563.

relationship depends upon the elastic value of Poisson’s ratio, the

shape of the stress-strain curve as given by the ratio of the secant to
the elastic modulus, and a plastic value of Poisson’s ratio.

Systematic experimental studies of the variation of Poisson’s ratio

h the yield region-are generally lacking in the literature. One study
is a report by Stang, Greenspn, and Newman (reference 2) for aluninum
alloys and low-carbon steels. Values of Poisson’s ratio under simple
tensile loading were obtained for strains as high as 18 percent on thin
flat tensile specimens.

It appears that a completely systematic series of tests should

include both tensile and compressive stress-strain properties along three
orthogonal axes as well as the Poisson ratio variation along these
directions under simple tensile and compressive loadings. An investi-
gation of this type was carried out for three comnonly used aluminum
alloys: Rolled 24s.T4 and extruded 14S-T6 and 75S-T6. The results are
given in the section entitled “Experimental Investigation.”

In the section entitled “Correlation of Theory and Test Data” the

validity of the theoretical relationship for the variation of Poisson’s .
ratio is examined by comparison with the experimental results.

This investigation was conducted under the sponsorship and with the .
financial assistance of the National Advisory Committee for Aeronautics.
The authors wish to acknowledge their indebtedness to Mr. Gary Gould for
valuable assistance in the experimental tivestigation and Mr. Conrad
Schmidt for machining of the test spectiens.

smBoI.s

E modulus of elasticity

E secant modulus

I quadratic strain invariant; defined in equation (6)

u normal stress

T shear stress

E normal strain a E
(t)
Y shear strain

v generalized Poisson’s ratio

(+)
-
 NACATN 2561 3

8 elastic strain (cJ/E)

5 strain deviation (C - C*
)
4 cubical dilatation ●

*
w variation of Poisson’s ratio from elastic value v -v )
( \
XYY)Z Cartesian coordinates

Superscripts:

* elastic co~nent

plastic component

Where two subscripts are used, the first refers to the direction
in which the load is applied and the second to the reference direction.

THEORETICAL CONSIDERATIONS

Transforms of Simple Tension to Simple Shear

..

For the purpse qf indicating the magnitude of the effect of the

Poisson ratio variation in a simple case, it is instructive to consider
the derivation of the affinity terms which transform a simple tensile
stress-strain curve into a simple shear stress-strain curve. Since
Poisson’s ratio is associated with strain only, the stress transformations
are written immediately.

According to the maximum-shear theory

T = o.~u (1)

For the octahedral-shear theory

The maximum-shear theory states that

(3)

In simple tension, v.s and thus from equation (3)

‘Y=- .
y=(l+v)~ (4)

. .. . -. . — .- —.—— —- —— -——- --————— - -- - - —- -— -— ... ---- .—--— . . .

 4

The octahedral shear strain can be written in the following form:

7oct = 0 81 3 (5)

where I is the quadratic strain invariant,

I 2 + # +Cz 2 +1~ 7xy2 + 7F2 + 7~ 2 (6)

,=i (‘x )( )

In simple tension

7oct = ‘M (7)

~ s@le shear

7oct = 7 ~/3
2 (8)

The affinity relationehip is obtained by equating equations (7)

and (8)

(9)

Thus, from equations (4) and (9), it is evident that the strain
affinity terms are actually functions of Poisson’s ratio. The lmliting
values which the affinity terms can assume for a plastically incompressi-
ble solid are given in the following table for typical values of Poissonts
ratio:

Lower limit ~Ver limit

5eory Affinity term @ = 0.3 = 0.5

Maximumshear (l+V) ●

Octahedral shear ~~j ;:;3 :

Theoretical J?oissonRatio Variation

me cubical dilatation of a strained solid is given by

$ = ~x + ~y+ ~z +

——. - --
 N/WATN 2561 5

If infinltes~ strains are-considered, which is a reasonable assump-

. tion in the yield region, the second- &d third’-rder terms of the -
dilatation may be neglected. ~h so doing, the dilatation is equal to
the linear strain invariant:

4 “x+ Gy+Ez (U)

The behavior of engineering materials indicates that the total

strain can be considered to be composed of two parts: The elastic strain
component .c* and the strain deviation 5. T5us, the dilatation can be
written

(
fl=ex*+6y*+6z*
)(+5x +5Y+5Z
) (12)

By considering the dilatation to be c’ompsed of elastic and plastic

components,

d =$*+T (13)

where

(15) ‘

The usual assumption of mathematical @asticity theory that the dilatation

vanishes is obtained by neglecting the elastic component and as$x
that the plastic component-is zero.

Consider a solid subjected to a simple tensile load in the x-

direction. Upon tiingthe assumption that the solid is isotropic along
the other two orthogonal sxes, the various dilatations can be determined.
Such a solid is referred to as orthot~pic.

For the elastic component, from equation (14)

O* = G*(1 - 2V*) (16)

and for the plastic component it is assumed that for the orthotropic
solid equation (15) can be written as

(17)
,

.- .- .. - . ...- --- _ _ -—.. -—- .- - --- .—.- -. —.- -—. . . -—— —. -. _.. — .——-——.. —
 I

6 NACATN 2561

The total dilatation, from equation (13) is

d = G*(1 - 2@) + 5(1 - 2~ (18)

It is further assumed that for an orthotropic solid the total dilatation

can be referred to the total strain by the relation

Combining equations (18) and (19) and simplifying, the variation of

Poissonis ratio as a function of strain is given by

(20)

For cases in which the plastic dilatation vanishes, ~= O, and

from equation (17), ~ = 0.5. In this s~eciai case, equation (20) reduces
to

v= 0.5 - $“0.5 -V*) .

or
.
v = 0.5 -EQ5 -V*) (20a)
$
When unloading follows loading into the plastic range, equations (20)
and (20a) yield the elastic value of Poisson’s ratio since the term
~*~ then becomes equal to unity.

An expression for the generalizejlPoisson’s ratio which corresponds

to equation (20a) is also given by N&Lai in reference 1.

EXPERIMENTAL INVESTIGATION

Description of Test Specimens

The materials under investigation were loaded in tension and com-

pression along each direction of a set of three orthogonal axes x, y,
and z where the z-direction is the direction of extrusion or rolling.
The materials tested were rolled 2kS-T4 aluminum alJoy and extruded I-4s-116
and 75S-T6 aluminum alloy. The spechens were cut from bars of square
cross section with sides equal.to # inches.
4

..-—— .—.—.— .—. . . ———... —— —.

 NACATN 2561_ 7

Engineering materials generally lack isotropy and, in addition, ‘the

properties may vary from point to point in the cross section. Therefore,
the specimens were cut from each bar in such a manner that strain measure-
ments were taken at essentially the same location in all tests.

A drawing of the tension and compression specimens is shown in’

figure 1. The tension specimens were loaded through special grips ‘
designed for this i.uvestigation. The grips were seated in spherical
bearings to insure application of axial loading to the specimen and are
shown with a specimen installed in figure 2. .

The compression specimens were mclxbed flat, square, and parallel

and carefully placed in the testing machine to minimize bending.

Test Procedure

Load was applied to the specimens by a Baldwti-Southwark universal

hydraulic testing machine of 200,~0-pound capacity with an accuracy of
loading of K1/2 percent. AX-5 strain gages were mounted on each of the
four sides of each specinEn. Awired tension specimen is shmn in figure 3.

Strains were measured with a Baldwin SR-4 strain tidicator. The

. estimate~ accuracy of the strain measurements is .approximatelyi2 percent.
The errors in the strain measurements are associated with the strain
indicator, the stated gage factor of th,estrain gage, and slight drift
of the strain readings at large plastic strains.

ltqerimental Results

A complete set of stress-strain curves for each of the aluminum

alloys tested is given in figures 4, 5, and 6. Poisson’s ratio h the
yield region is given in figures 7 to 12 for both tension and compression
with the load applied along each of the three coordinate axes. Poisson~s
ratio was computed by taking the negative of the ratio of corrected
transverse strain to strain in the direction of loading. Corrections
for the measured transverse strains are necessary because of the construc-
tion of the wire resistance strain gage. The method of correction is
given in the appendix.

A quantity of considerable interest in the theoretic&1 study was the

nature of the plastic dilatation. Accordingly, the plastic dilatation
as a function of strain deviation is given in figures 13, 14, and 15 for
the materials studied. The strain deviations were computed according
to the definition of this quantity, and the plastic dilatation was
obtained by equation (15).
8 NACATN 2561

Discussion of Experimental Results

In all cases, the values of strain sham in the figures were

obtained by averaghg each-back-to-back set of strain gages to eliminate
any bending. h the worst case, it was found that the msximum bending
strain was approximately 4 yercent of the axial strain.

In several of the plots of Poisson’s ratio variation shown in fig-

ures 7 to 12 it can be observed that the elaatic value of Poisson’s
ratio is not constant. Any scatter which exists at the first few loading
pints may possibly be attributed to expertiental technique. However,
the consistent variation of the elastic Poisson’s ratio shown for the
tension specimens in figure 10 suggests that a nonconstsmt elastic
Poisson’s ratio may actuallybe a property of the materials tested.
That such behavior was not observed in reference 2, and possibly in other
investigations, is attributed to the fact that the elastic Poissonts
ratio was computed as the ratio of the slopes of straight lines drawn
through stress+ial-strain and stress-trswmerse-strain data in the
elastic region.
.

.
CORREIATIOI? OF THEORYAItD TEST DATA
%
A significant feature of the experimental study is contained in
figures 13, 14, and 15 which show that the plastic dilatation is not
zero for the aluiinum alloys under hsvestigation. This experimental
fact maybe attributed, in part, to the anisotropic clmracter of these
engineering mterials. ,

The theoretical variation of Poisson’s ratio given in the section

entitled “Theoretical Consideration” was derived for an orthotropic
solid. The stress-strain characteristics of’the alloys used for the
expertiental investigation indicate that the cases given in the following
table maybe considered orthotropic if it is assumed that the transverse
strains are induced by an effective transverse load of the same sense as
the applied load. -

Material Loading
[

I 14S-T6
I Tension in z-direction

24s-T4 Tension in z-direction

2kS-T4 Compression in z-direction

I 75s-6 I Tension i-uz-direction

-———— ---- .- -.—.— ..-

RN rum m 2561 9
.

Thus, experimental data are provided for examination of the theo-

retical variation of Poisson’s ratio given in “Theoretical Considerations”{
by the relat”ion

v = T.qv. v’) (20)

Poisson’s ratio for various plastic strains maybe computed by use

of equation (20) if the following quantities are tiown:

(a) The elastic value of Ppisson’s ratio v * ,

(b) The stress-strain characteristicsof the material in the direc-

tion of application of load from which c*/c can be computed

(c) The term ~ which for a plastically incompressible isotropic

solid hhs the constant value of 1/2

The coefficient ‘~ bears further discussion since it is a term

which apparently incorporates the effects of nonvanishing plastic
dilatation. From equation (20) this coefficient may be expressed as

kiln v =V
G*/G-+o (a)
..

It is the asymtotic value of the Poisson ratio~variation to which the

elastic properties make no contribution, or it can be ,imaginedas the
strain ratio of a material in which the dilatations were purely plastic.
It is referred to hereafter as the asymptotic strain ratio.

Furthermore, the plastic dilatation for an orthotropic material was

assumed to be given by
.
3=5(1-27) (17)

If the asymptotic strain ratio is’s co~tant, then the plastic dilatation
should be a linear function of~the strain deviation.

An examination of fi~es 13, 14, and 15 for the orthotropic cases

listed in the preceding table indicated that although considerable
scatter does exist among the ‘experimentalpoints, a possible linear
relationship exists between plastio dilatation and strain deviation .
within the range of strains considered. A ’straightline was passed
through the test d&ta by th &methodof least squares and ~ was computed
from

.— .-. . . . .. .. .— ..- . ..— —— .- —- —.—.. .— .-—.— .——.—..——— -..

10 NACATN 2561

(l’i’a)

The values of V obtained from figures 13, 14, and 15 by means of

equation (17a) are given in the following table:

Material I Iaading Computed ~ 7-P

~4s-T6 Tension in z-dtiection 0.65 0.30

24s-T4 Tension in z-direction .56 .22

2kS-T.k Compression in z-direction .63 .29

ns-’l% Tension in z-direction .60 .25

To test further the validity of the theory by use of the ,experimental

data contained herein, it is proposed to examine these data by use of .
equation (20) to determine if the value of the asymptotic strati ratio
is constant for various strains and compares with the values given in
the preceding table. For this purpose, equation (20) can be rewritten
in the form
*
p=v ‘-V*ET ( 22)
1 -—~c*

Equation (22) canbe simplified by letting

where v is the change in v from the elastic value. With this

substitution, equation (22) reduces to.

(24)

For the cases listed in the table above, mlues of p as a ~ction

of ~ were obtatied from curves faired through the test data given in
figures 7, 9, and Il. The values of e*/e were computed from the stress- -
strain curves of the materials given in figures 4, 5, and 6. Then by use
of equation (24) the asymptotic strati ratios were calculated for the
. ,

-— ---
NACAm 2561. IL

series of selected points. These data are shown in figure 16 as com-

pared with the values of the asymptotic strain ratio given in the
preceding table.

Reasonably good agreement exists between the two methods of compu.

tation among all,the cases except for 14S-T6 aluminum alloy. The greatest
discrepancies between th two sets of data occur at values of e*/e
approaching unity. The nature of equation (24) is such @at large errors
are associated with this range of e*/6 values and consequently more
weight should be given to the points rewved from this region.

DISCUSSION

The Poisson~s ratio test data presented herein ‘forthe yield region
of the stress-strain curve exhibit the same general characteristics as
the test data given in reference 2 for similar aluminm alloys. The
tests reprted in reference 2, howeier, also go far beyond the @eld
region and indicate that in many cases the variation of Poisson’s ratio
reaches a maximum at a strain of between 2 and’6 percent and then
steadily decreases.

It appears that this behavior can be attributed to the fact ’that

beyond a strain of approximately 2 percent the stratis are no longer
small in the sense that the second- and third-order terms in the dila-
tation equation (10) may be neglected. This is demonstrated h refer-
ence 2 by computimg Poisson’s ratio for a plastically incompressible
isotropic solid using the following expression which can be derived from
equation (10) with the above assumptions:

,=M 6.
(3)

A comparison of nmrical results obtained from equation (a) and

the variation of Poissonts ratio given by equation (20) reveals that
the latter is adequate up to a strain of approximately 2 percent. Beyond “
this value of strain, equation (20) asymptotically approaches a maximum,
whereas equation (Z5) reaches a maximum and then decreases h substantial
agreement with the experimental behavior observed in reference 2.

The test data on plastic dilatation obtained from strain measure-

ments and shown in figures 13, 14, and 15 indicate that tensile loading
in the yield region is accompanied by a permanent decrease in volume,
whereas compression results in a permanent increase in volume. These
data were subsequently checked by density measurements on several
.

..-— -. —.. — ..——— ——— - ---- — .—— --— . -— --- ------ -

 12 NACATN 2561

specimens used in the experimental investigation of Poisson’s ratio and

also by independent volume measurements on a block of 14S-T6 aluminum
alloy compressed in the y-direction to various values of strain devia-
tion. From the latter, the data indicate that the volume changes noted
appear to be associated with the yield region only and apparently decrease
to negligible vol~e changes beyond a strain deviation of the order of
0.02 inch per inch.

CONCLUDING REMARKS

A systematic set of test data for the variation of Poisson’s ratio

in the yield region of the stress-strain curve is presented for the
aluminum alloys 14S-T6, 24s-T4, and 75S-T6. The test data are for simple
tensile or compressive loading along three orthogonal sxes.

For an orthotropic solid, a theoretical variation of Poisson’s ratio

in the @eld region was obtained from d.ilatationalconsiderations.
Certain of the test data indicated that under the loading used the
material could be considered orthotropic. These data were used to con-
firm the validity of the assumptions made in deriving the theoretical
variation of Poisson’s ratio.

Daniel Guggenheim School of Aeronautics

College of Engineering
New York University
New York, N.Y., November 14, 1950

— -. .,
 .
13

APPENDIX
.

GAGE-FACTOR CORRECTION FOR A OEE-DIMENSIONAL STRESS FIELD

The use of 90°.crossed-type (x) resistance strain gages in a one-

dlmensional stress field requires that a correction be applied to the
stated gage factor of the gage which is ~erpendicular to the applied
load. This correction arises from the fact that the end loops of the
transverse gage are subjected to a different strain from that used to
detemdne the stated gage factor. The stated gage factor of the strain
gage alined h the direction of the applied load requires no correction
since the manner of loading used in the test corresponds to that used
in calibration of the gage. “

The symbols used in the following discussion are:

R resistance of strain-sensitivewire

P resisitivity

L length of strain-sensitivewire
.
A area of strain-sensitivewire

k cmstant involving changes in p, L, and A

1 length of strain-gage grid r

1 w width of-strain-gage grid

n number of grip loops plus 1

,
,
G gage factor

r strain reading

The initial resistance of a strain gage in which the strain-

sensitive wire is arranged in the conventional rectangular grid’is

R. = POL A. (Al)
4
Under deformation, the resistance is changed by the,increment A(pL~A)
which involves changes,in p, . L, and A. It is known, however, that
this increment is a function of the deformation of the wire only. Thus

. . .. .. . ----- .-. -....-. .— ..-. .. ~.- .. . .- —-. —-- — — -----——- -. --- -—---— ----- --- --—--- -- - -—
 14 NACATN 2561
.

where is the strain along the axis of the gage and 62 is the
‘1
transverse strain. For loadings in which c . -VG2, equation (A2)
1
can be rewritten in the form

(A3)

where

The left-hand term of equation (A3) is the reading obtained from the
strain indicator. Equation (A3) can be written approximately as

(A4)

The corrected transverse strain is then

61= r -(1 - V2)62

(w/n2) (A5)

For the AX-5 strain gage used in the test described in the section
entitled “Experimental Investigation,” n = 10 and W/z = 0.5. ‘~US
.

‘l=r - O*4+2 (A6)

The only difficulty encountered in using equation (A6) to correct the

transverse strain measurements is that Poisson’s ratio was unknown.
As a first approximation V was calculated by taking the negative of
r/e~. It was found that a second approximation was unnecessary.

—.— .—--. —-—————— —. —-. -. —.-. .. —. ——— — -.-—- ----— --—-- --- . . —,
 NACAm 2561 15

Fm?ERENcEs
.

1. I&i, Ar_&d: Theory of Flow and Fracture of Solids. Vol. I.

Second ed., McGraw-~ Eoo.kCo., hC ., 1950, pp. 379-387.

2. Stang, Ambrose H., Greenspan, Martin, and Newman, Sanford B. :

Poisson’s Ratio of Some Structural Alloys for Iarge Strains. Res.
Paper RP 1742, Jour. Res,, Nat. Bur. of Standards, vol. 37, no. 4,
Cct. 1946, pp. 211-221.

. . . . ... —r . . . .. —.. — --—. —

—._. —.. — ..— ——- --. —. —-— ---—– -
 Tension specimen

Section A-A

l-%.-l

Compression sp8cimen

El< -11- I-14 ‘

Figure 1.- Tension and compression specimens used in experimental
investigation. V-ends are to be flat, squsxe, and parallel
within +0.002 inch.

— .— —.
——— ——————
 3N

,. 17

. .
. .---- . .. ..
---------- - ._
---. +
‘~ -.-— —- ...._
—-—-------- ~— —---...._--—
 IWCATN 2561. 19

t
Figure 5.- Stress-strain curves for”2@-T4 aluminum alloy.

-.. .— ..-. -.-. —-- --—- --------- -—-—- --—— -- —--- —--- —-— -——. — ...— — .— ----–-
 20 NACATN 2561
.

Figure 6.- Stress-strain curves for 75s-T6 ~uminum alloy.

-—.—. -. .
i

i
I

Figwe 7.- Poi6son’s ratio variation for 149-T6 aluminum alloy

loaded in z-direction.
1
+
I

I
1

I
I

—.
aGin, ax m uy, ia. /in.

Figure 8.- Poisson’s ratio variation for 14S-T6 aluminum alloy loaded
in x- or y-direction.

.— . .-
1

‘prdn ez, in. / *.

.—.

Figure 9.- Poisson’s ratto variation for 24s-T4” ELl~m EIIOy

loaded in z-direction.
 N
-F

“kl+aili,- ax-”oi “6-

~ Lb

Figure 10. - Poisson’s ratio variation for 24s-T4 alumlnum alloy

10*a in x- or y-direction.
 -’2

strain, GE, h/in.

Ftgure 11.- Poisson’s ratio variation for 75s-T6 aluminm alloy

loaded in z-direction.

.
I

CJ
P
Figure 12. - Poisson’s ratio variation for 75S-T6 aluminum alloy
loaded h x- or y-direction.
 5N NACATN 2561 27

Figure 13. - Plastic dilatation against strain deviation for

14s-T6 ‘aluminum alloy.

-—.. _ _ ___ _ .—— .—___ .. . ....

_— .——— —

28
.

.
Figure 14.- Plastic dilatation against strain deviation for
24s-T4 d.mlilllml.~Oy.

.— — —. _-. . .. ...— . ..-_ .——.

 6N NACATN 2561

Figure 15. - Plastic dilatation against strain deviation for

75S-T6 aluminum alloy.

..— ..-. .—-— —..-. .—_.. .._. —.__— ——. ,.—. .— -._..—. ... . . . . . . . ____________
 30 NACAm 2561
.

Figure I_6.- Comparison of computed values of v- V*.

NACA-1.m@6y-1-29-52- !0(,(,

. _— ...-.— ...—
—— -— —.